Top-level Fields

This document introduces the complete definitions of top-level fields in action.yml, including the configuration rules for name, version, author, description, inputs, outputs, and runs.

name

Required The name of the plugin. AtomGit Pipeline displays name in the pipeline orchestration interface, helping users intuitively identify each plugin in the steps.

name: 'codecheck'

version

Required The version number of the plugin, in semantic version format X.Y.Z.

version: '1.0.0'

author

Required The identifier of the plugin author, such as an employee ID.

author: 'XXX'

description

Required A one-sentence description of the plugin's functionality.

description: 'Sample plugin'

inputs

Required Input parameters. Input parameters can be used to specify data that the plugin uses at runtime. AtomGit Pipeline stores input parameters as environment variables. It is recommended to use lowercase input IDs.

When specifying inputs, the system creates an environment variable named INPUT_<VARIABLE_NAME>. The input name will be converted to uppercase and spaces will be replaced with _ characters.

inputs:
  key_input:
    description: 'Single-line input test'
    required: true
    default: test

Note If no input is specified, an operation with required: true will not automatically return an error; it needs to be validated explicitly in the code.

inputs.<input_id>

Required A string identifier associated with the input. The value of <input_id> is a mapping of input metadata. <input_id> must be a unique identifier in the inputs object, must start with a letter or _, and can only contain alphanumeric characters, -, or _.

inputs.<input_id>.description

Required A string description of the input parameter.

inputs.<input_id>.required

Optional A boolean indicating whether the operation requires the input parameter. Set it to true if the parameter is needed.

inputs.<input_id>.default

Optional A string representing the default value. This is used when the input parameter is not specified in the pipeline orchestration file.

outputs

Required Output parameters. Output parameters can be used to declare data set by the plugin, and subsequent operations in the workflow can use the output data from previous operations.

outputs:
  record_id:
    description: "recordId"

outputs.<output_id>

Required A string identifier associated with the output. <output_id> must be a unique identifier in the outputs object, must start with a letter or _, and can only contain alphanumeric characters, -, or _.

outputs.<output_id>.description

Required A description of the output parameter.

runs

Required Specifies the execution method and runtime of the current Action.

runs.using

Required The execution method of the Action. The following three options are supported:

Execution Method	Description
node16	Node.js version, executes JavaScript/TypeScript compiled code

runs.main

Required The JavaScript code entry file corresponding to the Action execution.

runs:
  using: 'node16'
  main: 'dist/main.js'

runs.post

Optional The JavaScript code entry file corresponding to the termination of the Action, used for cleaning up, executing post-processing logic, etc.

runs:
  using: 'node16'
  main: 'dist/main.js'
  post: 'dist/stop.js'

Trigger mechanism for post:

Trigger Method	Description
Manual Stop	User clicks to stop the pipeline, called actively by the scheduling service
Natural Call	Automatically called after the plugin runs (the plugin developer needs to actively listen for termination signals in main and call the post logic)

Example - Listening for termination signals in main:

// Import the relevant post method
import {post} from "./stop";

async function run() {
    // Capture the SIGINT signal
    process.on('SIGINT', () => {
        // Call the post logic
        post();
        process.exit(0);
    });

    // Plugin main logic...
}